Low-dose CT image denoising using sparse 3d transformation with probabilistic non-local means for clinical applications

Abstract

The use of low-dose computed tomography (LDCT) in medical practice can effectively reduce the radiation risk of patients, but it may increase noise and artefacts, which can compromise diagnostic information. In this paper, we propose a LDCT denoising method based on sparse 3d transformation with probabilistic non-local means (PNLM). The hard-thresholding module in the sparse 3d transformation framework is used to attenuate noise in the transform coefficients. In addition, the PNLM overcomes the incompetence of non-local means weights problems where its weights better reflect the patch similarities, thereupon it connects the denoising process and the noise type and thus able to denoise complex noise present in LDCT images. Besides, a significant denoising improvement is obtained by using the collaborative wiener filtering. Experiments on NIH-AAPM Mayo Clinic LDCT dataset show that the proposed method can effectively remove noise while recovering finer details and provide better visual perception than other state-of-the-art methods.